🤖 DiabetWise AI Process Flow
Intelligent Food Detection & Diabetes Risk Assessment
🔄 AI Detection Process
📸
Capture Image
User takes photo of food using smartphone camera
→
🔍
AI Processing
Dual AI models analyze food type and nutritional content
→
⚠️
Risk Assessment
Algorithm determines diabetes safety level
→
💡
Smart Recommendations
Personalized advice and portion suggestions
🧠 Dual AI Architecture
🍽️ Food Detection AI
YOLOv8 object detection
EfficientNet classification
2,000+ Indonesian food database
90%+ accuracy target
Traditional food specialization
Real-time processing
📊 Calorie Estimation AI
ResNet CNN architecture
Volume/portion estimation
±15% accuracy target
Depth detection algorithm
Nutritional value calculation
Macro/micronutrient analysis
🚦 Smart Warning System
🟢
SAFE
Low carbs, low glycemic index. Safe for diabetics to consume in normal portions.
🟡
CAUTION
Moderate carbs/sugar. Consume in smaller portions with blood sugar monitoring.
🔴
DANGER
High sugar/carbs, high glycemic index. Avoid or consult doctor before consuming.
📈 Indonesia Diabetes Statistics
10.7M
Diabetics in Indonesia
73%
Undiagnosed Cases
60%
Diet-Related Complications
$2.3B
Annual Healthcare Cost
📚 Validated Data Sources
TKPI 2020
Indonesian Food Composition Table
USDA FoodData
Central Nutrition Database
Food-101
Food Image Recognition Dataset
UEC-Food100
Asian Food Specialization
Custom Dataset
Indonesian Traditional Foods
🛠️ Technology Stack
📱
Flutter
🧠
TensorFlow
🔥
Firebase
⚡
FastAPI
🗄️
PostgreSQL
🚀
Redis
🌟 Expected Impact
👥
500K+
Users in First Year
💰
$150M
Healthcare Cost Savings
📉
30%
Complication Reduction
🏥
100+
Puskesmas Integration
📚
60%
Awareness Improvement
⏱️
2 sec
Analysis Speed